AbstractIn this paper, we give arguments that there are two types of coronal mass ejection (CME).The first type of CME discussed here is the ‘loop-type’, whose occurrence is related to an arcade flare somewhere between the footpoints. It was found that there were pre-event magnetic connections between the flare location and the locations of the footpoints of a CME of this type, and that these connections disappeared after the event. This suggests that the footpoints of loop-type CMEs are special prescribed points, and this was verified by the observation that the footpoints do not move in this type of CME.The other type of CME is the ‘bubble-type’, which is associated with the flare blast from explosive flares. We confirmed the association of this type of CME with the so-called EIT (Extreme Ultra-violet Imaging Telescope) waves, but the velocity of expansion of the bubble is twice or more greater than that of the EIT waves depending on events. Although EIT waves were widely considered to be Moreton waves viewed by SoHO/EIT in the solar activity minimum period, recent simultaneous observations of both have revealed that the EIT wave is something different from the Moreton wave, and propagates separately with a velocity less than half that of a Moreton wave.We therefore propose a new overall picture: the bubble-type CMEs are the flare-produced MHD blast waves themselves, whose skirt is identified as a Moreton wave. EIT waves may be interpreted as follows: the slow-mode gas motions from the source cause secondary longwavelength fast-mode waves which are trapped in the “waveguide” in the low corona. The secondary long-wavelength wave in the fast-mode, which is trapped in the low corona, has a slower propagation velocity due to the nature of the waves trapped in a “waveguide”. This trapped wave induces slow-mode motions of the gas through a mode-coupling process in the high chromosphere, where the propagation velocities of the fastand slow-mode waves match.Three-dimensional MHD simulations for these two types of CME are in progress, and are previewed in this paper.
Genetic slow-mode/fast-mode optimisation of digital PID controllers
